EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Energy-Based Stability Control for Suppressing Memory Poisoning in Multi-Agent Collaborative Systems

Lukas Schneider, Anna Becker, Felix Wagner

Fundamental Scientific Reports in Multidisciplinary Areas · 01.04.2026 · 2 (1) · 24-31

DOI: 10.66238/fsrma51 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Fundamental Scientific Reports in Multidisciplinary Areas, online erschienen am 01.04.2026. Verfasst von Lukas Schneider und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Memory poisoning can destabilize collaborative agent dynamics by introducing divergent update trajectories. This study models system stability using a Lyapunov-inspired energy function defined over collective memory states. An adaptive damping mechanism constrains updates that increase system energy beyond predefined bounds. The method was tested on a 150-agent cooperative optimization benchmark across 5,000 iterations. Under a 20% poisoning scenario, baseline systems exhibited energy divergence in 64.5% of runs. Energy-controlled systems reduced divergence probability to 17.8% and maintained 91.6% of nominal task efficiency. Energy-based stability control offers a control-theoretic approach to suppress contamination-induced instability.

Zitieren

APA

Schneider, L., Becker, A., & Wagner, F. (2026). Energy-Based Stability Control for Suppressing Memory Poisoning in Multi-Agent Collaborative Systems. Fundamental Scientific Reports in Multidisciplinary Areas, 2(1), 24-31. https://doi.org/10.66238/fsrma51

Vancouver

Schneider L, Becker A, Wagner F. Energy-Based Stability Control for Suppressing Memory Poisoning in Multi-Agent Collaborative Systems. Fundamental Scientific Reports in Multidisciplinary Areas. 2026;2(1):24-31. doi:10.66238/fsrma51

BibTeX
@article{Schneider2026,
  author = {Schneider, Lukas and Becker, Anna and Wagner, Felix},
  title = {Energy-Based Stability Control for Suppressing Memory Poisoning in Multi-Agent Collaborative Systems},
  journal = {Fundamental Scientific Reports in Multidisciplinary Areas},
  year = {2026},
  volume = {2},
  number = {1},
  pages = {24--31},
  doi = {10.66238/fsrma51}
}
Kostenlos starten →